Having unveiled in April what it dubs a business user interface for Hadoop, startup AtScale today announced a $7m funding injection that it says will help it bring the technology to market.
The series A investment round, led by UMC Capital, with AME Ventures, Storm Ventures and XSeed Capital, will help finance an expansion of the San Mateo, California-based firm's sales, marketing and customer-support activities.
AtScale's Intelligence Platform is an application designed to allow firms to use their existing business-intelligence tools to access data in Apache Hadoop, acting as an interface to the Hadoop ecosystem and SQL-on-Hadoop engines such as Impala, Spark and Spark SQL, Hive on Tez, Presto and Drill.
"Right now, we're really confident that the product has all the features that we need. Of course, we'll continue to expand it but we have the team in place already to be able to do that. We're not anticipating a big growth in the engineering side, mostly in sales, marketing and customer support," AtScale founder and CEO David Mariani said.
AtScale, which was set up by staff who worked at Yahoo during the early days of Hadoop, is aiming to invest both in technical support as well as training.
"Hadoop is still very new, especially to the business-intelligence customer. There's a big learning curve, not just in, 'How do you do this new style of business intelligence that we put out?' but also it's about, 'How do I set up a Hadoop cluster?' and 'What are the pitfalls and what do I need to do to avoid failure?'," Mariani said.
Rather than working exclusively with firms with Hadoop early-adopter expertise, AtScale has in fact mainly been dealing with companies with no Hadoop experience and which use business intelligence extensively but are running into the scale limitations that Hadoop could solve.
"They're usually kicking the tyres of Hadoop. But we're introducing the idea that they can truly move their business-intelligence workloads to Hadoop, which is surprising them in such a way that, 'Yes, let's get moving on Hadoop. Let's get beyond the tyre-kicking stage and really look at putting this into production'," he said.
Such companies are also finding it difficult to secure the service they need from the main Hadoop distribution firms.
"The big distributions are so busy with their own resources when it comes to consulting and the like, a lot of the business-intelligence customers can't get their attention," Mariani said.
"So they look to us for advice on how to structure their files for business intelligence, what SQL on Hadoop engine they should use, Should it be Impala? Should it be Spark SQL? Should it be Hive Tez? How should they structure those files and how should they migrate from their traditional OLAP of BI infrastructures to Hadoop. That's where we add a lot of value."
AtScale's Intelligence Platform application makes Hadoop appear to business users as a multi-dimensional OLAP cube, in which data is already classed into dimensions, measures and hierarchies to speed up queries.
The subscription-based software is installed on a client or gateway node on a Hadoop cluster, rather than on the cluster itself and does not require either a separate cluster, dataset or set of servers to manage the cube.
Investors in this round include also Amr Awadallah, founder and CTO at major Hadoop distributor Cloudera. Michael Franklin, director at Berkeley UC AMPLab, the origin of Apache Spark and cluster manager Mesos among other technologies, is a company adviser.